/*
 * Licensed to the Apache Software Foundation (ASF) under one or more
 * contributor license agreements.  See the NOTICE file distributed with
 * this work for additional information regarding copyright ownership.
 * The ASF licenses this file to You under the Apache License, Version 2.0
 * (the "License"); you may not use this file except in compliance with
 * the License.  You may obtain a copy of the License at
 * 
 *      http://www.apache.org/licenses/LICENSE-2.0
 * 
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */
package org.apache.forrest.util;

import java.io.IOException;
import java.io.Serializable;
import java.util.Map;
import java.util.Stack;

import org.apache.avalon.framework.parameters.Parameters;
import org.apache.avalon.framework.service.ServiceManager;
import org.apache.avalon.framework.service.ServiceException;
import org.apache.cocoon.ProcessingException;
import org.apache.cocoon.transformation.AbstractDOMTransformer;
import org.apache.cocoon.caching.CacheableProcessingComponent;
import org.apache.cocoon.environment.SourceResolver;
import org.apache.cocoon.util.HashUtil;
import org.apache.excalibur.source.SourceValidity;
import org.apache.excalibur.source.impl.validity.NOPValidity;
import org.apache.excalibur.xml.dom.DOMParser;
import org.apache.excalibur.xml.xpath.XPathProcessor;
import org.w3c.dom.Attr;
import org.w3c.dom.Document;
import org.w3c.dom.Element;
import org.w3c.dom.Node;
import org.w3c.dom.NodeList;
import org.xml.sax.SAXException;

/**
 * A transformer that prunes the source tree based on <code>include</code> and
 * <code>exclude</code> XPath expressions.
 * The <code>include</code> parameter specifies nodes to let through, and
 * <code>exclude</code> parameter nodes to filter out. Either or both may be
 * omitted. The default behaviour is for all nodes to be let through
 * (<code>include</code> = "/").
 * <p>
 * This transformer should be declared in the sitemap at
 * <b>map:sitemap/map:components/map:transformers</b>, as follows<br>
 * <pre>
 * &lt;map:transformer logger="sitemap.transformer.xpath" name="xpath" src="org.apache.cocoon.transformation.XPathTransformer" /&gt;<br>
 * </pre>
 * <h3>Example usage</h3>
 * As an example, consider a user manual XML file:
 * <pre>
 * &lt;manual&gt;
 *   &lt;s1 title="Introduction"&gt;
 *     &lt;p&gt;This is the introduction&lt;/p&gt;
 *     &lt;p&gt;A second paragraph&lt;/p&gt;
 *   &lt;/s1&gt;
 *   &lt;s1 title="Getting started"&gt;
 *     &lt;p&gt;Getting started&lt;/p&gt;
 *     &lt;fixme&gt;Add some content here&lt;/fixme&gt;
 *     &lt;n:note xmlns:n="urn:notes"&gt;banana 1&lt;/n:note&gt; &lt;note&gt;banana 2&lt;/note&gt;
 *     &lt;p&gt;Yes, we have no bananas&lt;/p&gt;
 *   &lt;/s1&gt;
 * &lt;/manual&gt;
 * </pre>
 * We could now deliver named chapters as follows:
 * <pre>
 *   &lt;map:match pattern="manual/*"&gt;
 *      &lt;map:generate src="manual.xml"/&gt;
 *      &lt;map:transform type="xpath"&gt;
 *        &lt;map:parameter name="include" value="/manual/s1[@title='{1}']"/&gt;
 *      &lt;/map:transform&gt;
 *     &lt;map:serialize type="xml"/&gt;
 *   &lt;/map:match&gt;
 * </pre>
 * So <code>manual/Introduction</code> would return the first chapter.
 */
public class XPathTransformer
    extends AbstractDOMTransformer
    implements CacheableProcessingComponent
{

    /** XPath Processor */
    private XPathProcessor processor = null;
    private DOMParser parser = null;

    /** XPath specifying nodes to include. Defaults to the root node */
    protected String include = null;
    /** XPath specifying nodes to exclude. Defaults to "" (no exclusions) */
    protected String exclude = null;

    public void setup(SourceResolver resolver, Map objectModel,
            String source, Parameters parameters)
        throws ProcessingException, SAXException, IOException {
        super.setup(resolver, objectModel, source, parameters);
        this.include = parameters.getParameter("include", "/");
        this.exclude = parameters.getParameter("exclude", null);
        }


    public void service(ServiceManager manager) throws ServiceException {
        super.service(manager);
        try {
            this.processor = (XPathProcessor)this.manager.lookup(XPathProcessor.ROLE);
        } catch (Exception e) {
            getLogger().error("cannot obtain XPathProcessor", e);
        }
        try {
            this.parser = (DOMParser)this.manager.lookup(DOMParser.ROLE);
        } catch (Exception e) {
            getLogger().error("cannot obtain DOMParser", e);
        }
    }


    /** Implementation of a template method declared in AbstractDOMTransformer.
     * @param doc DOM of XML received by the transformer
     * @return A pared-down DOM.
     */
    protected Document transform(Document doc) {
        getLogger().debug("Transforming with include='"+include+"', exclude='"+exclude+"'");
        Document newDoc = null;
		try {
        	newDoc = handleIncludes(doc, this.include);
        	newDoc = handleExcludes(newDoc, this.exclude);
		} catch (SAXException se) {
			// Really ought to be able to propagate these to caller
			getLogger().error("Error when transforming XML", se);
			throw new RuntimeException("Error transforming XML. See error log for details: "+se);
		}
        return newDoc;
    }

    /**
     * Construct a new DOM containing nodes matched by <code>include</code> XPath expression.
     * @param doc Original DOM
     * @param xpath XPath include expression
     * @return DOM containing nodes from <code>doc</code> matched by <code>XPath</code>
     */
    private Document handleIncludes(Document doc, String xpath) throws SAXException {
        if (xpath == null || xpath.equals("/")) {
            return doc;
        }
        Document newDoc = parser.createDocument();
        NodeList nodes = processor.selectNodeList(doc, xpath);
        for (int i=0; i<nodes.getLength(); i++) {
            Node node = nodes.item(i);
            addNode(newDoc, node);
        }
        return newDoc;
    }

    /**
     * Construct a new DOM excluding nodes matched by <code>exclude</code> XPath expression.
     * @param doc Original DOM
     * @param xpath XPath exclude expression
     * @return DOM containing nodes from <code>doc</code>, excluding those
     * matched by <code>XPath</code>
     */
    private Document handleExcludes(Document doc, String xpath) {
        if (xpath == null || xpath.trim().equals("")) {
            return doc;
        }
        NodeList nodes = processor.selectNodeList(doc, xpath);
        for (int i = 0; i < nodes.getLength(); i++) {
            Node node = nodes.item(i);
            // Detach this node. Attr nodes need to be handled specially
            if (node.getNodeType() == Node.ATTRIBUTE_NODE) {
                Attr attrNode = (Attr)node;
                Element parent = attrNode.getOwnerElement();
                parent.removeAttributeNode(attrNode);
            } else {
                Node parent = node.getParentNode();
                if (parent != null) {
                    parent.removeChild(node);
                }
            }
        }
        return doc;
    }

    /**
     * Add a node to the Document, including all of the node's ancestor nodes.
     * Eg, if <code>node</code> is the title="Introduction" node in the
     * example, the doc would have node /manual/s1/@title='Introduction' added
     * to it.
     *
     * @fixme This method could do with some optimization. Currently, every
     * node addition results in one expensive node equality check per ancestor
     * @param doc Document to add a node to
     * @param nodeTemplate Node from another Document which we wish to
     * replicate in <code>doc</code>. This is used as a template, not actually
     * physically copied.
     */
    private void addNode(Document doc, final Node nodeTemplate) {
        // Get a stack of node's ancestors (inclusive)
        Stack stack = new Stack();
        Node parent = nodeTemplate;
        Document oldDoc = nodeTemplate.getOwnerDocument();
        while (parent != oldDoc) {
            stack.push(parent);
            parent = parent.getParentNode();
        }
        // Example stack: (top) [ /manual, /manual/s1, /manual/s1/@title ] (bottom)

        // Now from the earliest (root) ancestor, add cloned nodes to the
        // doc. We check if a suitable ancestor node doesn't already exist in
        // addNode()
        parent = doc;
        while (!stack.empty()) {
            Node oldNode = (Node)stack.pop();
            Node newNode = null;
            if (!stack.empty()) {
              // Shallow copy o a parent node (in example: /manual, then /manual/s1)
              newNode = doc.importNode(oldNode, false); // Do a shallow copy
              copyNamespaceDeclarations(oldNode, newNode);
              parent = findOrCreateNode(parent, newNode);
            } else {
              // Deep copy of the matched node (in example: /manual/s1/@title)
              newNode = doc.importNode(oldNode, true);
              copyNamespaceDeclarations(oldNode, newNode);
              parent.appendChild(newNode);
            }
        }
    }

    /**
     * Add xmlns namespace declaration attribute to newNode, based on those from oldNode.
     * It seems that a DOM object built from SAX with namespace-prefixes=false
     * doesn't have xmlns attribute declarations by default, so we must
     * manually add them.
     * @param oldNode Original node, with namespace attributes intact
     * @param newNode If an Element, this node will have an <code>xmlns</code>
     * (or <code>xmlns:prefix</code>) attribute added to define the node's namespace.
     */
    private void copyNamespaceDeclarations(final Node oldNode, Node newNode) {
      if (newNode.getNodeType() == Node.ELEMENT_NODE) {
        String prefix = oldNode.getPrefix();
        String nsURI = oldNode.getNamespaceURI();
        Element newElem = (Element)newNode;
        if (nsURI != null) {
          if (prefix == null || prefix.equals("")) {
            if (!newElem.hasAttribute("xmlns")) newElem.setAttribute("xmlns", nsURI);
          } else {
            if (!newElem.hasAttribute("xmlns:"+prefix)) newElem.setAttribute("xmlns:"+prefix, nsURI);
          }
        }
      }
    }
 
    /**
     * Add newNode as a child of parent, first checking if any equivalent node
     * to newNode already exists as a child of parent.
     *
     * @param parent Parent node of found or created node
     * @param newNode The node potentially added to parent, unless parent
     * already has an equivalent node
     * @return the appended node, or the old equivalent node if found.
     */
    private Node findOrCreateNode(Node parent, Node newNode) {
        NodeList otherChildren = parent.getChildNodes();
        for (int i = 0; i < otherChildren.getLength(); i++) {
            Node child = otherChildren.item(i);
            if (nodeEquality(child, newNode)) {
                // Found existing equivalent node
                return child;
            }
        }
        // No existing equivalent node found; add and return newNode
        parent.appendChild(newNode);
        return newNode;
    }

    /**
     * Shallow-test two nodes for equality.
     * To quote from the Xerces DOM3 Node.isEqualNode() javadocs, from where
     * most of the code is filched:
     *
     * [Nodes are equal if] the following string attributes are equal:
     * <code>nodeName</code>, <code>localName</code>,
     * <code>namespaceURI</code>, <code>prefix</code>, <code>nodeValue</code>.
     * This is: they are both <code>null</code>, or they have the same length
     * and are character for character identical.
     */
    private boolean nodeEquality(final Node n1, final Node n2) {
        if (n1.getNodeType() != n2.getNodeType()) {
            return false;
        }
        if (n1.getNodeName() == null) {
            if (n2.getNodeName() != null) {
                return false;
            }
        }
        else if (!n1.getNodeName().equals(n2.getNodeName())) {
            return false;
        }

        if (n1.getLocalName() == null) {
            if (n2.getLocalName() != null) {
                return false;
            }
        }
        else if (!n1.getLocalName().equals(n2.getLocalName())) {
            return false;
        }

        if (n1.getNamespaceURI() == null) {
            if (n2.getNamespaceURI() != null) {
                return false;
            }
        }
        else if (!n1.getNamespaceURI().equals(n2.getNamespaceURI())) {
            return false;
        }

        if (n1.getPrefix() == null) {
            if (n2.getPrefix() != null) {
                return false;
            }
        }
        else if (!n1.getPrefix().equals(n2.getPrefix())) {
            return false;
        }

        if (n1.getNodeValue() == null) {
            if (n2.getNodeValue() != null) {
                return false;
            }
        }
        else if (!n1.getNodeValue().equals(n2.getNodeValue())) {
            return false;
        }
        return true;
    }

    // Unused debugging methods

    private final void printNode(String msg, Node node) {
        getLogger().info(msg+" "+node.getNodeName());
    }
/*
    private final void printDeepNode(String msg, Node node) {
        try {
            getLogger().info(msg+" "+XMLUtils.serializeNodeToXML(node));
        } catch (ProcessingException pe) {
            getLogger().error("Error printing node", pe);
        }
    }
*/
    // Cache methods

    /**
     * Generate the unique key.
     * This key must be unique inside the space of this component.
     *
     * @return A hash of the include and exclude parameters, thus uniquely
     * identifying this XPathTransformer amongst it's peers.
     */
    public Serializable getKey() {
        return ""+HashUtil.hash(this.include+this.exclude);
    }
    // For backwards-compat
    public Serializable generateKey() {
      return getKey();
    }

    /**
     * Generate the validity object.
     *
     * @return An "always valid" SourceValidity object. This transformer has no
     * inputs other than the incoming SAX events.
     */
    public SourceValidity getValidity() {
        return new NOPValidity();
    }

    // for backwards-compat
    public SourceValidity generateValidity() {
      return getValidity();
    }

    /**
     * Recycle the component.
     */
    public void recycle() {
        super.recycle();
        this.include = null;
        this.exclude = null;
        // note that we don't turf our parser and processor,
    }

    /**
     * dispose
     */
    public void dispose() {
        super.dispose();
        this.processor = null;
        this.parser = null;
        this.include = null;
        this.exclude = null;
    }
}